If We Don't Believe Women, AI Won't Either: Using A Feminist Innovation Lens To Build A Brighter Future For All
By Raji Mohanam, Founder and Editor in Chief, Women in the AI Vanguard Magazine
We are building the most powerful decision-making technology in human history inside systems that still struggle to decide whether to believe a woman.
This fall, a former Cornell student filed a lawsuit alleging she was drugged and sexually assaulted by members of a campus fraternity in 2024, and that the university, the fraternity, and her former sorority failed to protect her. Within days, New York's governor called for an independent review of whether Cornell met its obligations to its students, and the university agreed to an outside investigation. Prosecutors have since reopened the criminal case. The details belong to her and to the courts, but the disappointing pattern belongs to all of us.
Most women have seen the pattern before, up close. It exists in all sectors. I've spent more than two decades in medical education, research, and public health, and I've watched it play out in real-time there. In American emergency rooms, women wait an average of thirty minutes longer than men for pain medication. Clinicians handed identical charts, with only the patient's gender changed, rate the woman's pain as less severe. A campus, a hospital, a boardroom: different spaces yet they all discount a woman's voice.
Now, not surprisingly, we're hardwiring that structure into code.
We need to hit pause. AI is moving at high velocity into the very places where credibility gets decided: the triage algorithm that ranks whose symptoms are urgent, the hiring model that sorts whose experience counts, the reporting and case-management tools institutions lean on when a woman comes forward. Each of these systems learns from the record our institutions have already written, and that record carries the discount. Train a model on decades of dismissed pain and quietly closed complaints, and it learns to dismiss and close, faster and at scale, with the authority of math behind it. The AI industry has an answer ready, and it calls that answer "ethical and responsible AI". I'd ask the harder question: responsible to whom, and decided by whom? An institution that can't hear a woman in the room will build a machine that can't hear her either. Truly responsible AI asks for a different way of seeing, one where the people most affected by a decision sit at the table when it's made and hold the power to stop it. That shift has to happen now, while the defaults are still being set. Once they're locked into code, they'll outlast every feeble task force convened to fix them.
An institution that can't hear a woman in the room will build a machine that can't hear her either.
That's why the lens through which our decision-makers see and solve problems matters more than the tools. A women-centered, inclusive way of seeing starts by asking who carries the risk before it asks what a product can do, and it treats the people most affected as co-authors of every decision. Tech companies need to adopt that lens now, for reasons this particular moment makes urgent. AI is moving from answering our questions to acting on our behalf, making choices inside hospitals, classrooms, courtrooms, and workplaces. The money, the talent, and the architecture are consolidating around a small circle of builders, and the norms they set this year become the infrastructure the rest of us inherit. Public trust is already fraying, and every institutional failure like the one unfolding at Cornell teaches people that power protects itself. But a company that can create solid foundations alongside the people most often discounted, earns a trust that no marketing budget can buy. Women-centered innovation is how ethical and responsible AI can finally become real.
WHAT IS A FEMINIST INNOVATION LENS?
The Feminist Innovation Lens is a way of building technology that starts with the people most often discounted. It treats women, and everyone the system has learned to overlook, as co-authors of innovation rather than edge cases to patch later. Drawn from matrilineal societies that have shared power for centuries, it holds that ethical innovation depends on three questions: who carries the risk, who shares the decision, and who can say stop. For institutions, it's a practical standard for responsible AI. For everyone else, it's a promise that the future gets built with women in the room.
Where Power Moves in a Circle
If the lens is the problem, it's fair to ask where a better one comes from. Women have been modeling it for centuries, in societies the rest of the world has rarely thought to learn from.
In West Sumatra, the Minangkabau have passed land, homes, and family names through women for generations. The great ancestral houses belong to the women of the lineage, and the men who grow up in them carry real roles as brothers, uncles, and advisors. Decisions are made through a practice the Minangkabau call musyawarah, deliberation that continues until the community reaches agreement. It's slow by design. Everyone affected gets heard before anything moves.
Across the world, in what is now upstate New York and southern Ontario, Haudenosaunee clan mothers have long chosen the chiefs who speak for their nations. They also hold the power to remove a chief who stops serving the people. Leadership is a trust, and the women of the clan hold its keys. In the hills of Meghalaya in northeast India, Khasi families name the youngest daughter as keeper of ancestral property, a custodian holding it for the generations who come after her. Among the Mosuo near Lugu Lake in southwestern China, households gather around a woman at the center who manages the family's shared resources, and decisions are made together.
These are living societies with their own histories, tensions, and pressures from a world that often misunderstands them. They are teachers rather than utopias. And what they teach is remarkably consistent.
Anthropologist Peggy Reeves Sanday spent years living with and learning from the Minangkabau. She came away arguing that "matriarchy" had been misread for a century. In her account, it describes a society where maternal values, nurturing, continuity, and shared responsibility, anchor the social order. Women stand at the center, with everyone held inside the circle.
Sanday knew the other side of the picture too. She also wrote Fraternity Gang Rape, her study of how brotherhood, privilege, and secrecy on American campuses produced sexual violence and then protected it. Earlier, comparing societies around the world, she found that sexual violence clustered where men dominated, interpersonal violence was common, and women's roles were devalued. Few scholars have looked so closely at both ends of the spectrum, and her work points in one direction: who holds power over decisions shapes who gets hurt and who gets heard. The young woman at Cornell and the women of a Minangkabau longhouse live at opposite ends of that same line.
German scholar Heide Göttner-Abendroth went on to map what these societies share. Economically, they circulate wealth to keep the community in balance. Socially, they organize kinship through the mother's line. Politically, they decide by consensus, starting in the household and building outward to the village and the region. Culturally, they treat the land, the ancestors, and the generations to come as part of the same community. She describes them as societies in balance. Power flows across the circle, and it answers to the people it touches.
This is where the Feminist Innovation Lens comes from: an inclusive way of seeing that treats ethical innovation as a question of who holds power. And it's why the lens asks three questions of any innovation:
Who carries the risk?
Who shares the decision?
Who can say stop?
In a Minangkabau longhouse or a Haudenosaunee council, those questions are ordinary. Inside most tech companies, they're radical. And they are exactly the questions this moment demands.
Who Gets Believed Decides What Gets Built
Asking those three questions changes the work from its very first step, the moment someone decides which problems AI will confront. Big tech has built a habit of aiming its most powerful tools at the problems easiest to see from inside its own walls: engagement, efficiency, ad performance, the speed of a workflow. Step outside that room and a different list appears. The 2 a.m. symptom a woman can't get anyone to take seriously. The complaint that sits in an inbox for months. The deepfake that travels faster than any takedown request. The unpaid hours of care holding up every household and every economy.
Women have been naming these problems for generations. The trouble has always been who gets believed. Philosopher Miranda Fricker calls it testimonial injustice: the credibility a speaker loses because of who she is. When the woman describing a problem is believed less, the problem itself becomes harder to see, and it slides off the list of things worth solving. The Feminist Innovation Lens corrects for that at the source. It gives her a seat in the room where the problems get chosen, so her account shapes the work before the work begins.
Centering women and other marginalized communities this way is a design decision that ends up serving everyone. When cities cut ramps into sidewalk curbs for wheelchair users, parents with strollers, travelers with suitcases, and delivery workers all moved more freely. Build AI that confronts the problems of the people who don't get heard, and you build AI that listens better to anyone the system has learned to overlook. Researcher Afsaneh Rigot built a whole methodology on this idea. Her Design From the Margins framework centers the most marginalized and impacted users from the earliest stage of design, and she argues the result is safer, better technology for everyone. That's inclusive design at its most practical.
This already works in practice. Earlier this year, we published an article by founder Emily Richardson who described how she builds the AI agents that run her company, KNWN. She noticed that nearly everyone building agents was copying the old corporate org chart into code, complete with a chain of command. She chose a heterarchy instead, a horizontal structure guided by matriarchal values. Her system, The Council, is made up of five AI stewards working side by side, guided by distributed authority, consent before action, and transparency. Her stewards carry no gender. The matriarchy in her system lives in its values and in how its parts communicate, and it runs a real business.
Consensus is what keeps that lens in place. One room, however brilliant, sees one set of problems. A table where the people affected share the decision, and can say stop, sees the rest. Richardson proved it can be built. The question now is whether the companies with the most power will choose to build it.
The Matriline Method: A Framework for Responsible AI
A lens only matters if it changes what we build. So for the organizations building and deploying AI that are ready to act, I'm offering a simple place to start. I call it the Matriline Method, named for the line of inheritance that has carried these societies through centuries of change. It turns the three questions into four steps.
Trace. Before you build, follow the line. Find where women live inside your data, your supply chain, and your blind spots. Ask whose labor trained the model, whose stories are missing from it, and who carries the risk when it gets something wrong.
Time. Give the people most affected the time to weigh in while decisions are still being made, and real standing when they do. Good decisions, as the Minangkabau know, are slow by design. That includes the power to slow a launch or stop it, the way a clan mother can remove a chief. A seat with no say is decoration.
Tend. Measure what care actually looks like in your product. Count the invisible work that keeps it running, from the data labelers to the trust-and-safety teams, and pay for it. Judge success by whether people end up better cared for.
Test. Test every decision against the world it leaves behind. Ask what your system will mean for the daughters who inherit it.
Every one of these steps asks for a decision far more than a breakthrough. The tools already exist. The research has been done. What the industry now needs is a different way of seeing, and leaders with the will to adopt it.
The young woman at Cornell, the patient waiting in the ER at 2 a.m., the daughter who will inherit whatever we build this year. They are the same thread, and the future of AI runs straight through them.
So I'll ask the leaders reading this: when your next product/model is ready to ship, who in your company has the power to stop it? And would a woman's objection be enough?
This essay opens a WAIV series on the Feminist Innovation Lens. Coming next: an in-depth guide to each step of the Matriline Method, with practical detail on how to put it to work across your organization. Join the Vanguard to get it first.
Raji Mohanam is the Founder and Editor of Women in the AI Vanguard Magazine, where she highlights the women and systems shaping ethical, responsible innovation that can meet human needs. She brings more than two decades of work in medical education, research, and public health to the question of who technology serves.
Sources and Further Reading
• Emily Richardson, "Rethinking AI: Building a Future with Heterarchical Values," WAIV Magazine, 2026.
Barbara Alice Mann, Iroquoian Women: The Gantowisas, Peter Lang, 2000